maps

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implementation uses only Python's standard library modules, avoiding the security risks associated with third-party package dependencies.
  • [SAFE]: Network activity is restricted to recognized mapping and location services (OpenStreetMap, OSRM, TimeAPI.io) for their intended purposes.
  • [SAFE]: Place names and geographic coordinates provided as user input are correctly URL-encoded before interaction with external APIs, mitigating injection risks.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data and user-provided inputs, creating a standard surface for indirect interaction.
  • Ingestion points: User-provided place names and addresses via CLI arguments; geographic data returned from Nominatim, Overpass, OSRM, and TimeAPI.io.
  • Boundary markers: None explicitly defined in the script logic.
  • Capability inventory: Network access (HTTP GET/POST) to specific mapping APIs.
  • Sanitization: All user inputs are sanitized via standard URL encoding libraries (urllib.parse) before being transmitted to external services.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 07:01 PM
Security Audit — agent-trust-hub — maps